Several years ago, when passing the house where my father grew up, I noted an odd distinction. Dad, it seemed, had been more familiar with the families that had lived on his street in Cincinnati than I had grown to be, a generation later, with those who lived near our house outside Buffalo. Friendly neighbors had animated his childhood in ways in which they were entirely absent from mine.

I might have left it at that. In the course of our day-to-day lives, we all have a tendency to note little things, and most commonly we set those observations aside. But this particular distinction stuck with me. I began wonder whether there might be a connection between the political dysfunction on display on Capitol Hill and the changed rhythms of American neighborhood life. Was what I’d observed typical of other American parents and their children? And how, if it existed, had that broader shift rippled across American society?

Exploring those questions in the journey that would eventually lead to my new book, The Vanishing Neighbor, I soon ran into a roadblock. The unfortunate truth is that there are no dispositive data to confirm or contradict the contention that America’s social architecture has changed. To be sure, bits and pieces of evidence offer hints. In Bowling Alone, Robert Putnam compiled convincing reams of data showing that the connective tissue that held previous generations of Americans together was in decay. In Diminished Democracy, Theda Skocpol similarly found that the nature of voluntary organizations was changing.

But there was no hard evidence to measure whether, as I had come to suspect, Americans were choosing to invest their limited funds of time and energy in different buckets of relationship. At the same time we were bowling alone, I suspected, we were doubling down on our most intimate relationships and expanding the breadth of superficial acquaintances. It wasn’t so much that we were becoming more or less connected, but that the connections themselves were beginning to shift.

In the end, I got lucky. Harvard University’s Peter Marsden pointed me to data from the General Social Survey, in which Americans reported on whom they had socialized with over the course of the previous month. More Americans were associating with family members and contacts who lived a few miles away, and fewer were spending time with their neighbors.

In the absence of that statistical evidence, the power of my observation might well have been viewed as nothing but conjecture. But should it have been? Should phenomena we observe without precise methods of measurement be made the subject of sustained scholarship?

A few generations ago, the answer to those questions would have been obvious. Many of the most consequential books of the 1940s and 50s—David Riesman’s The Lonely Crowd and William Whyte’s The Organization Man—were powerful not because they unearthed new data about American society so much as because they observed changes in the patterns of everyday American life. Each argued that the American character was changing, as individuals put greater emphasis on the need to fit in with their peers. There was barely a statistical footnote in either book. But while some observers would come to dispute the authors’ conclusions in the wake of the social upheavals of the 1960s and 70s, it wasn’t just the ideas that were dismissed—it was the method of scholarship.

Over the past several decades, the value we once put in simple observations has been eclipsed by the power of statistics. It’s not that anyone believes that data should be ignored. In many cases, hard figures offer us valuable insights and empower us to hold individuals and organizations to account. How much more dire is the problem of economic inequality? Which charitable strategies are best equipped to help the most people? Particularly in the world of philanthropy, metrics are a crucial bulwark against the risk of wasting precious resources.

But it’s come now to the point where hard data aren’t just the means to a conclusion—data have emerged as the parameter for our questions. Not long ago, I was talking with a friend who was working toward a Ph.D. Discussing the prospect of his dissertation, he explained that he wasn’t sure what the topic would be, because he “hadn’t yet found any good data.” In essence, rather than determining what deserved in-depth scholarship, he was sizing up the available data and asking what might be gleaned.

A little more than two years ago, I flew out to Berkeley to meet with the legendary (and now late) sociologist Robert Bellah. In the course of our conversation, I voiced my concern that I didn’t have sufficient hard evidence to prove my theory. Bellah gave me a knowing smile, sighed, and recalled a story from earlier in his career. He had been out to lunch with the anthropologist Clifford Geertz and a few other scholars when a topic of some significance came up in conversation. Someone interjected, “Well, we really don’t know whether that’s right, because we have no studies on that.” To which Geertz retorted, “Well, you live in the society and have eyes, don’t you?”

The point isn’t that we ought not to be rigorous in testing and challenging suppositions. It’s not that we should ignore hard facts when they’re available. But the world of scholarship—whether inside the academy or out—can’t be limited to the boundaries of measurable data. There are too many important questions for which we can’t compute answers. The absence of numerical evidence shouldn’t discourage an investigation. Quite the opposite: If a question is worth answering, then the underlying issue should be considered worthy of simple and sustained observation.

Author Bio: Marc J. Dunkelman is a research fellow at Brown University’s A. Alfred Taubman Center for Public Policy and American Institutions. His first book, The Vanishing Neighbor: The Transformation of American Community, was published this month by W.W. Norton.